Throughout the last few years, Robotics Process Automation (RPA) was touted as a universal cure for many shared services and BPO branch vulnerabilities. It was billed as a cheap platform that was easy to deploy, scale and deliver, and required no special knowledge to get things done. Vendors propagated their marketing message well enough that enterprises believed their statements. On top of that, every “new tech” phenomenon allures with its promises, and RPA was no exception. RPA also created a new political tool and an additional row in Excel spreadsheets that managers could use to play with results.
“Cheap,” “affordable” and “cost neutral” are tricky yet appealing words inside shared services and BPO, which are already called “cost centres” for a reason: each year they try to slice any possible cost by a significant margin. It’s quite rare to see them eagerly invest in and allocate budget to a new product that has not been battle-tested. Yet with RPA, they did. Management, driven by sales pitches and a few sample-based POCs, decided to include in their projections the 30% cost savings RPA promised to achieve.
Bringing RPA back to reality
Unfortunately, this risk was overplayed. When the reality hit the fan, people realised that the expected results from RPA were mainly driven by very synthetic models that were isolated — believe or not — from the quite complicated delivery models and processes inside shared services. People also realised that RPA is only effective when the data is clean and structured. Unfortunately, the process of capturing, organising, validating and delivering unstructured data in a structured form to a digital workforce is expensive and time-consuming. In fact, the cost of converting unstructured data to structured data very easily nullifies the ROI from RPA.
So what can we conclude – that everybody was fooled by insidious marketing and after all that struggle we are left with nothing?
Outstanding transformation through RPA
Fortunately, the outlook is much more promising. From the program management chair where I have been watching, the transformation progress impelled by RPA is truly outstanding.
RPA forced organisations to finally map their processes 100%. Companies found that the “lean” process study methodology used for the last decade is absolutely not enough to enable a deep transformation journey. RPA demands that engineers diagnose, analyse and diagram all variations, permutations, exception handling, judgement points, IT dependencies and voice of the customer touch-points. There is no room to miss even a single event or dependency.
As a result, we have witnessed a growing number of service providers creating dedicated products for process mining, including Celonis Process Mining and DXC Agile Process Automation (APA).
RPA demands new employee skills and the creation of completely new employee positions. Employees need to be agile like never before. Not only do they need an analytical mindset, but also the ability to create robotic workflows and to logically combine sets of actions. New “transformation experts and specialists” need to understand more than simple Java, .NET and C++ coding. Their supervisors must learn a hybrid organisation style to manage an environment where a portion of the work is done by robots and the remainder by humans.
The environment also necessitates new management of change models, where operations accommodate requests from bots and accept human performance evaluations by robotic creations. Oh, and did I mention that it all requires new ways to calculate results and new designs for metrics? Yes, it does, and it is exciting too.
It is also very interesting to see new ways of cooperation between old-school IT managers/app owners and operations robotics managers. The robotics revolution has pushed groups to work together and to penetrate the closed castles inside their kingdoms — without hurt egos. Due to direct app connections, massive volumes pushed in seconds, security, accesses, servers, disks and much more, robotics simply cannot be kept outside the IT circle of trust.
Furthermore, as the cost of overall support has grown significantly throughout the last two decades, IT has come to respect robotics. Managers know one of the ways to make ITO services slimmer is to enable robotics and to automate. Such an amazing synergy!
Process mining is not the only stream to benefit from the robotics revolution. Soon after the first POCs and RPA projects were delivered, managers speculated that to really transform freely and not be limited by small restrictions or data permutations, RPA needed to become a bit more intelligent, spot patterns, leverage an increasing number of AI algorithms and enable machine learning.
As a buyer, you probably saw that shortly after the first RPA software wave, some vendors refocused on Smart Process Automation (SPA) or re-positioned their software as Intelligent or AI driven. This forced transition was caused by vendors themselves falsely presuming that people in back office operations spend all of their time following strict rules and performing simple processing of structured streams of data. The RPA revolution definitely influenced vendors in this area to search, innovate and advance towards being even more universal and agile.
Can we let it fly?
To close this fascinating thread, my take is that robotics started in a sexy, appealing, yet really premature form, with huge promises that people decided to trust and bet their money on. Surprisingly, when the fight was almost lost, nobody decided to give up. Vendors started to re-build RPA and make it more useful. Enterprises — with their employees, organisations and mechanisms finally ready for change — decided to give it another shot.
The result? RPA is not going away. It was just a caterpillar transforming into butterfly. The transition was hard, maybe even painful for some, but currently it is up to us to decide if we will allow it to fly.